330 / ОтКомпанииTI / doc7559Стеклопод

AVR480: Anti-Pinch System for Electrical Window

Automotive motorized window or door apparatus which closes automatically involve risks for trapping, squeezing or injury to people. They must reverse their movement in case the force applied by the motor gets higher than one normalized limit. The feature implies the constant monitoring of the speed, current, position of the glass. For cost and simplicity reasons, the system described in this application note uses universal brush motor equipped with hall effect sensors. The detection algorithm is based on speed and torque derivate and has been verified for robustness and fault tolerance. It is applicable to all ATMEL AVR® with A to D converter and Interrupt-on-Change I/O.

Microcontrollers

ApplicationNote

7559A–AVR–06/06

Powered

Equipments in

Modern Cars

The Standards

Today, electronic components and systems account for over 20% of the cost of a high end passenger car. The ability of dense electronics to better control sensors and actuators is utilized to enhance comfort and safety in cars and it is very predictable that most of the middle and upper class vehicles will be systematically equipped with motorized window or door systems.

The majority of these equipments come with full automation, means that they must be accompanied by safety systems to prevent people injury or mechanical degradations.

The legislation offers set of rules to which powered systems must comply. This is particularly true for window lifts or sliding doors. This application note describes how to implement an antipinch algorithm which was initially developed for a powered window but which can be easily adapted to any other moving part.

The automotive power-operated windows are governed by international standards, like MVSS118 in USA or 74/60/EEC in Europe. For those relatives to the elimination of danger to children, the requirements expressed in these documents can be summarized as follows (see Figure 1):

•Detection area: 4 mm to 200 mm,

•Maximum pinch force = 100N,

•Reverse direction on a pinch.

•Tests with determined deflection ratio objects: 5N/mm to 20N/mm

Figure 1. Anti-Pinch window lift rules as per MVSS118

4mm

200mm

Hardware

Consideration

Anti-Pinch

Algorithm

Specification

Solution Using

Motor’s Physical

Parameters

Modeling,

Simulation

AVR480

Different detection strategies are possible to determine whether an obstacle is entering the critical pinching area:

•Without mechanical contact. It reacts before the pinch effort is exerted on the object. This is the optimal protection since no force is applied to the obstacle. It is also independent to vibration, aerodynamic variations, deformations. But, it requires integrated sensors (infrared, ultrasonic, ...) with their electronic modules and wires leading to additional costs.

•With mechanical contact. The pressure measurement will tell the system an object is being pinched. Also there, designers have two fundamental technologies available:

–Direct measurements: Force sensors or contactors are integrated on the door seal. These solutions are inherently high cost and reduce the styling for window/door designs

–Indirect measurements via physical monitoring. This is a globally cost optimized solution.

The pinch detection algorithm shall, at first, respect the standards (FMVSS118 & 74/60/EEC) requirements:

•Detection area from 4 to 200mm

•Maximum exerted force of 100N

•Reverse direction on Pinch detection

•Normalized tests for validation

It has to be adaptive since:

•The mechanical elements involved in the lift system will vary with time (ageing, local deformation, wear, ...)

•The electrical characteristics could evolve significantly

•The environment will affect the friction forces (temperature, moisture, freeze, etc.)

The system should not react to disturbances neither detect inopportune pinches. It must be

robust against air friction, road vibration, power-cuts, etc...

The force applied to the glass can be extrapolated from the current through the motor. The position of the moving elements permanently provides information on speed. Both parameters can then be used to determine whether an obstacle has been encountered and whether:

•It is in the detection area

•The force applied gets higher than the limit

This document describes one anti-pinch algorithm developed to work with motor current measurements and hall-effect speed indications. With very little changes, it can adapat to other systems like slidding doors or roof panels.

To better design the pinch-detection algorithm, the physical parameters involved in the lift system have been extracted. Starting from a generic description of an opening/closing aparatus, one more precise model has been developed for the powered window. It is broken down into several main components as illustrated in Figure 2.

3

Motor Model

The identification of the mechanical components leads to consider:

•Motor Effects

–Mechanical Parameters

–Electrical Parameters

•Window-Lift:

–Frictions (limited to slides)

–Machanical non-linearity (coupling mostly)

–Inertia (glass weight)

The DC universal motor used can be modeled using a very classical scheme as illustrated in Figure 3.

Figure 3. DC Motor Equivalent Model

Figure 4. Second Order Motor Model

E

K

C1

M

C2

C1 ≈ nr (M&z&2 + Fv z&2 + Fs + Mg )

F

Z

F1

Z1

Z2

F2

Weight

Pinch Detection

Algorithm

Pinching Condition

Usual pinch-detection algorithm operation is using indirect measurements out of the window-lift system:

•Current (torque)

•Position (speed)

The algorithm detailed in this document agregates two techniques based on

•Calibrated torque stored in non-volatile memory: Preliminary learning sequence is executed and torque values are stored in memory. This is quite memory consuming and requires regular calibration sequences

•Speed derivate calculation: Interesting method since it requires less memory but needs more computing power

and takes benefit from the two methodologies.

A pinch is detected by comparing the current measurement with a reference (see Figure 7).

M(t) ) kϕ x i(t) = K x i(t)

The threshold can be determined out of the Motor Torque constant (K) combined with the response time of the system (Motor and the load).

Figure 7. Condition for Pinch

Other Interesting

Functions

Figure 11. Downwward Movement Finite State Machine

Timeout

Bottom position Or Timeout

Calibrate / Init of the position

Not described here are secondary functions necessary for Window-lift operation:

•Measuring current, filtering